Software Engineer - Automatic Computer Use (Windows)

Role Overview

We are seeking a Windows-focused Software Engineer to build agent-driven automatic computer use capabilities that can observe, reason, and take actions on Windows PCs. This role combines deep Windows system engineering with LLM-powered agent integration, enabling intelligent, safe, and automated computer interactions.

You will work on software agents that leverage LLMs for reasoning and decision-making, while executing actions through Windows OS APIs and services.

Key Responsibilities:

Windows Computer Use & Automation

• Design and implement agent actions that interact with Windows OS, applications, and system services
• Build reliable computer use primitives such as application control, system configuration, file operations, and window management
• Integrate with Windows APIs, services, and background tasks
• Ensure agents function correctly across user sessions, power states, and device conditions

LLM Integration & Agent Intelligence

• Integrate LLMs into agent workflows for task planning, decision-making, and intent interpretation
• Implement prompt templates, tool-calling patterns, and structured outputs
• Handle fallbacks between LLM-based reasoning and deterministic logic
• Optimize LLM usage for latency, cost, and reliability

Agent Execution & Reliability

• Implement execution logic including retries, rollback, and error handling
• Optimize system performance, power usage, and responsiveness
• Support long-running and event-driven agent workflows
• Security, Privacy & Trust
• Respect Windows security boundaries and permission models
• Implement safeguards to ensure transparent and user-controlled automation
• Contribute to logging, auditing, and explainability of agent actions

Collaboration

• Partner closely with Agent Platform, AI, UX, and Android teams
• Participate in architecture reviews and cross-platform design discussions
• Help define best practices for AI-driven computer use automation

Basic Qualifications

• 4+ years of professional software engineering experience
• Strong experience building Windows desktop or system-level software
• Proficiency in C++ and/or C# (.NET)
• Hands-on experience with Windows Services, background execution, or system APIs
• At least 2 years of experience integrating and using Large Language Models (LLMs) in production or near-production systems, including:
  • Calling LLM APIs or frameworks
  • Integrating LLMs into application workflows
  • Handling structured outputs, error cases, and latency considerations
• Strong understanding of multithreading, concurrency, and system debugging

Preferred Qualifications

• Experience with agent-based automation or computer use systems
• Advanced LLM integration experience, including:
  • Tool / function calling
  • Prompt engineering and evaluation
  • Hybrid rule-based + LLM architectures
• Knowledge of Win32, WinRT, UWP, WMI, or PowerShell
• Familiarity with:
  • Power management and resource constraints
  • Enterprise or OEM Windows environments
• Exposure to AI safety, guardrails, or explainability techniques